The high-resolution map of Oxia Planum, Mars; the landing site of the ExoMars Rosalind Franklin rover mission

This 1:30,000 scale geological map describes Oxia Planum, Mars, the landing site for the ExoMars Rosalind Franklin rover mission. The map represents our current understanding of bedrock units and their relationships prior to Rosalind Franklin’s exploration of this location. The map details 15 bedrock units organised into 6 groups and 7 textural and surficial units. The bedrock units were identified using visible and near-infrared remote sensing datasets. The objectives of this map are (i) to identify where the most astrobiologically relevant rocks are likely to be found, (ii) to show where hypotheses about their geological context (within Oxia Planum and in the wider geological history of Mars) can be tested, (iii) to inform both the long-term (hundreds of metres to ∼ 1 km) and the short-term (tens of metres) activity planning for rover exploration, and (iv) to allow the samples analysed by the rover to be interpreted within their regional geological context.


Introduction
Oxia Planum is the landing site for the European Space Agency's (ESA) ExoMars rover mission (Vago et al., 2018).The missions goal is to search for signs of past and present life.To accomplish this, the ExoMars rover, called Rosalind Franklin, will investigate the geochemical environment in the shallow subsurfacethe rover is equipped with a drill to retrieve samples from as deep as 2 musing its 'Pasteur' suite of scientific instruments during Rosalind Franklins nominal surface mission of 218 sols (Vago et al., 2017).
In anticipation for the surface mission, the ExoMars Rover Science Operations Working Group (RSOWG) established a 'Macro' sub-group dedicated to preparations for operations relating to orbital and rover remote sensing data analyses.The rover can drive 30-100 m per day, depending on the terrain's complexity, so it is important to (1) characterise the landing site with metre-scale fidelityrelevant for day to day rover operations,and (2) to cover an area encompassing the uncertainty ellipse for likely landing locations (which has an area in the order of ∼1000 km2 ).The search for signs of life by the rover is predicated on identifying the best possible sites for assessing three parameters: (1) geological context consistent with life-hospitable conditions, (2) potential physical biosignatures, and (3) potential chemical biosignatures.
To achieve this, the RSOWG Macro group organised a two-phase, high-resolution geologic mapping group campaign (Sefton-Nash et al., 2020) focused on the 'one-sigma' landing ellipse (i.e. an ∼66.75 × 5 km ellipse with ∼67% touchdown probability) with in Oxia Planum.The campaign consisted of an initial 'group mapping' phase, where team members mapped one or more 1 km square 'quads', followed by a 'reconciliation phase'.In the group mapping phase, the principal objectives were (i) to familiarise scientists with the geography and geology of the landing site, and (ii) to collect observations to create a geological map to build and guide the mission's strategic planning.The end goal is making the best use of the available mission resources (e.g.time, energy, locomotion range, data) to identify locations for sample acquisition and analysis.In the reconciliation phase, the phase 1 observations were reassessed to create a consistent map description of the geology.The reconciliation phase incorporated interpretations from elsewhere in the 3-sigma landing ellipse (i.e. the ∼ 100 × 20 km ellipse with 99% touchdown probability) and beyond.
The wider goals of this project are (i) to develop a thorough understanding of the landing site's geography and geological history prior to operations, (ii) to formulate site-specific testable hypotheses to facilitate interpretation of mission results, and (iii) to map locations where those hypotheses could be tested.This map will therefore be vital for both contextualising the scientific analysis of the rover samples.Here we present the completed map including descriptions and the stratigraphic relationship of map units and discuss how its intended use for rover operations, planning and future scientific study.

Location and projection
The main map covers an area from 17.11°N to 19.15°N and from 334.60°E to 336.60°E (Figure 1).The map is presented at 1:30,000, with digital products are available as vector and raster layers with a vertex spacing of 4 m and pixel size of 1 m in an equirectangular projection centred at 18.20°N, 335.45°E, based on the IAU Mars2000 sphere (available through Fawdon et al., 2021).This matches the coordinate system used by the ExoMars Rover Operations Control Centre (ROCC) to minimise local distortion.This is important for maintaining accuracy between remote sensing observation and rover scale operations.

High resolution imaging science experiment (HiRISE)
The map was created from a base-map mosaic (Volat & Quantan-Nataf, 2020; Figure 2) of red-channel HiR-ISE (McEwen et al., 2007) orthoimages and digital elevation models (DEMs) created using HiRISE stereo pairs following the method of Kirk et al. (2008) detailed in Fawdon et al. (2021) and Volat and Quantan-Nataf (2020).The HiRISE mosaic was georeferenced to a 6 m/pixel Context Camera (CTX; Malin, 2007) DEM and orthoimage mosaic (Fawdon et al., 2021) which in turn has been georeferenced to the 12 m/pixel High-Resolution Stereo Camera (HRSC;Gwinner et al., 2016;Neukum & Jaumann, 2004) MC11W regional mosaic (DLR, 2019) and MOLA spheroid (Neumann et al., 2003;Smith et al., 2001).The HiRISE mosaic was used to control Phase 1 mapping line work and all other data used were georeferenced to it.This ensures metre-scale spatial consistency between the various datasets.Where available, individual HiRISE Stereo DEMs, ortho images, Infrared, Red, Blue (IRB) and synthetic colour (RGB) data were used on top of the Volat et al. (2022) mosaic to provide additional information.These data used in Phase 2 reconciliation were georeferenced to the orthorectified red-channel image mosaic and are listed in Table 1 and any further information will be made available on request to the corresponding author.

Colour and stereo surface imaging system (CaSSIS)
The CaSSIS (Thomas et al., 2017) camera has four filters (IR; 950 nm, NIR; 850 nm, PAN; 650 nm and BLUE-GREEN; 475 nm), which provide a multispectral capability sensitive to a variety of minerals (Tornabene et al., 2017).We used all available CaSSIS data (Table 2) for the Phase 2 Oxia Planum work, but found the Near, Pan, Blue (NPB) band combination to be particularly useful.NPB images were georeferenced as described in Fawdon et al. (2021).

Making the map
The two-phase approach that we applied is similar to that used in geological assessment of the Mars 2020 landing site (Stack et al., 2020).The mapping focused on the one-sigma landing ellipse and covered the range of ellipse azimuths (the arrival azimuth rotates depending on the launch date within the approximately two-week-long launch window).This relatively small area was chosen in an effort to balance the human resources required to map at a level of fidelity relevant to the scale of rover operations (characterising variation in geology over tens of metres) and the infeasibility of the whole 3-sigma pattern of landing ellipses.In the event that the rover lands outside our mapped region, a rapid reconnaissance mapping (RRM; Balme et al., 2019) protocol is being developed in which a dedicated team will quickly map the terrain in a small region around the landing location.RRM will use the geological framework already established in this work to extend our understanding to the new location.This approach is possible because the 'one sigma' ellipse region is geologically representative of the wider area within Oxia Planum (Fawdon et al., 2021).

Group mapping
In Phase 1 group mapping, 75 volunteers (noted as 'Mapping team' in the author list) associated with the RSOWG (first 19 authors in the author list) followed a programme of four training sessions on mapping practice and contextual geology, four area discovery meetings on designated mapping areas, and a pre-mapping meeting in June 2020.Then, before mapping started, mappers were presented with a 'unit guide' based on preliminary observations of the main units and features.The exercise organisers also provided the mappers with the mapping symbology and conventions to be used.The quad grid and naming conventions for the geography of Oxia Planum were also established in this phase and reported in Fawdon et al. (2021).1), HiRISE colour images (Table 2) and (d) CaSSIS NPB mosaic (Table 3) used in Phase 2. (e) Completed Quads from the Phase 1 group mapping prior to reconciliation and (f) showing the Phase 2 final reconciliation output.
The 75 volunteers were then assigned quads to map at a fixed scale of 1:2000.Mapping was completed in the NASA/JPL authored open-source Multi-Mission Geographic Information System (MMGIS; Calef et al., 2017;Calef & Soliman, 2019) with additional functionality for geospatial analysis and mapping, 'CAmpaign Mapping and Planning' (CAMP) developed for the NASA Mars 2020 rover mission.The ESA ExoMars team uploaded the required data and configurations necessary to facilitate group mapping to MMGIS/CAMP.The base-map datasets used for this Phase were the HiRISE orthoimage and DEM mosaics (Volat & Quantan-Nataf, 2020) and the CaS-SIS colour mosaic (Fawdon et al., 2021).Between June 2020 and August 2020, the mapping volunteers concurrently mapped geologic contacts and surface features in 2D and with a supplementary 3D projection of the data to each complete one or more quads and wrote a 'quad abstract' recording their observations (Figure 2).

Reconciliation
In this phase, a small team (the two lead authors, in conjunction with advice from the lead authors) reconciled the scientific observations and linework made by the individual mappers to create a unified set of interpretations and a final map product.Reconciliation consisted of two steps.
In the first step, data ingestion, the geological units, contact and feature linework, and points of interest  recorded by the volunteer mappers were ingested into ArcGIS Pro (v.2.9-3.1)GIS software along with new HiRISE DEM and orthoimages (Table 1), new CaSSIS colour data (Table 3), CRISM and OMEGA hyperspectral data summary products (see: Brossier et al., 2022;Carter et al., 2015Carter et al., , 2023;;Mandon et al., 2021;Parkes-Bowen et al., 2022;Quantin-Nataf et al., 2021) describing the distribution of hydrated minerals to support reconciliation.In addition, certain classes from the machine-learning terrain classification maps of the NOAH-H (Novelty Or Anomaly Hunter-HiRISE) project (Barrett et al., 2022(Barrett et al., , 2023) ) were used, mainly those representing aeolian bedforms and boulder fields.
In the second, scientific decision-making step, the insights from the volunteer mappers, recorded in their 'Quad abstracts' were fused with new observations and new contacts digitised at a fidelity appropriate for rover operations.To achieve this input from the mappers was as a 'field slip' line work redrawn to a consistent scale and style by the lead authors to create the reconciled 'fair copy' map (e.g.; Barnes & Lisle, 2013).The map digitisation was conducted in Arc Pro in 2D alongside concurrent 3D local scenes to guid the placement of unit boundaries.Mapping following US Geological Survey planetary mapping conventions (e.g.Skinner et al., 2022) adapted to meet the aims of the study at a scale of 1:2000.The geology of Oxia Planum is highly variable over short (tens of metres) scales but homogenous over long (kilometre or greater) scales.Units were challenging to discriminate and map at a large enough scale to be relevant to rover operations because of the textural and tonal heterogeneity over short, 10's m, length scale compared to the homogeneity over length scales of 1 km to 10 of km.In addition to the bedrock groups, summary overlays of surficial features and units were created based upon the NOAH-H results, constrained by new human observations.These mapping decisions are discussed in Section 4.1.

Visualisation choices
The selection of colours and symbology decisions were made to enable the ease of communication across the diverse international science team.Bedrock groups where coloured to intuitively follow their appearance in CaSSIS data with group that covered small areas of the map (E.g.; DM, BM and MM) given high contrast hues to aid identification whilst respecting reader sensitivity (such as colour blindness).Impact related material is yellow to follow the global conventions with red added as the sequential end member denoting the most recent events.Within each group the strongest hue denotes the recent unit as per (Skinner et al., 2022).Three line styles denote our confidence in the existence and location of each unit boundary.These range from 'inferred'where logically a boundary should be located, through 'uncertain'where we can tell a boundary is present, but it is not directly observed to 'certain'where a contact is directly observed.Surficial units are denoted with shades of black and grey with no separate boundary symbology and textural overlays are denoted by black patterned areas.For familiarity global elevation uses a 'MOLA like' colour scales whilst a divergent scale in muted tones is used locally to highlight topographic differences.

Measurements
When the map sheet was complete, cross sections and summary scientific publications were produced.To aid the cross-section and unit descriptions several measurement of 'unit thickness have been made.These measurements have been made using the mean difference in 10 pairs of spot hights collected at the shortest path across the unit margin whist considering the quality of the underlying DTM and should be only considered estimates of the apparent thickness.

Results
The main map (CTX DEM and ORI mosaics) (Figure 3) covers 8750 km 2 , spans an elevations range of ∼ 600 m, and from −2992 m in the Aquitania region to the SE to −3156 m at the base of a crater floor in thePannonia region to the NW the mean eleevation is -3077 m with a standerd devation of 35 m.The map consists of 15 bedrock units organised into 6 groups, 7 textural and surficial units (see Description Of Map Units (DOMU): Table 4) and structural interpretations.The bedrock units represent geological processes dating back to the mid-Noachian.Our interpretation of the stratigraphic relationships, and their relative timing relative to other regional aqueous, erosional, and tectonic processes is shown in the Correlation Of Map Units (COMU) diagram (Figure 3).These relationships and their significance will be detailed in a future publication on the Geological history of Oxia Planum.

The discrimination between units and key stratigraphic relationships
This map is a detailed investigation into geological units in Oxia Planum that will support exploration with the Rosalind Franklin rover.The spatial constraints of the study area, dictated by the mapping scale required to fulfil the objective of the map, relative to the size of geological features in the region mean that the map cannot cover a sufficiently large extension to formulate complete hypotheses.To address this challenge, we have defined units with enough high fidelity to aid exploration and be representative of the wider Oxia Planum region.Here we outline these decisions in stratigraphic order and discuss open questions presenting avenues for further exploration.

Bedrock units
In the lower bedrock group (lBg) we map three units, all strongly associated with spectral signatures of aqueous alteration.We chose to be generous in retaining division between the units in the lower bedrock group to identify where differences in the expression of the Figure 3.The map of rock units and correlation of map units.The 1:150,000 scale map of bedrock units from the main map and the correlation of map units (Skinner et al., 2022).The timing of events is drawn from the contextual geological literature and the timing of other events in the region (e.g.Sections 1.2 & 4.2).Coloured bars represent the time in which these events are likely to have occurred, and not representative to their duration.We have not derived impact crater size frequency distribution model ages of for map units because, following the discussion in Sun and Stack (2020), these techniques is not an appropriate technique for units of this size and in this complex state of erosion/exhumation (e.g.Warner et al., 2015).There is some association with ripples (surface can be 'wavy').No resolvable compositional associations.
Predominantly occur in association with uBg1 and lBg1.Crater rims removed and little or no ejecta remains.No resolvable compositional associations.Can be superposed by S, Mm and uBg units.
Randomly distributed and not related to specific geologic unit.Forms wide (>100 m) upstanding areas in all regions of the map.Strongly associated with the 'honeycomb' texture and scarps in Pannonia.Some association with phyllosilicate minerals and possible mixtures with olivine (Mandon et al., 2021).Dark regolith distribution below the CRISM pixel size may obscure this signal.

Dark group oDm
Overlies or grades downwards into uBg3, low hills of this unit are often capped with uBg1.Directly underlies uBg1 in mesa walls.Occurs thoughout the Oxia Planum region.Crater wall cross-sections show sedimentary architecture.
Small patches of S t and D material frequently overlie the surface complicating identification.
Not sampled in any CRISM data.
Occurring as a layer around the bottom of larger mesas of uBg2 and appears to underly uBg2.Appearing as outliers overlying lBg units.Evident in the stumps at the base of rounded buttes capped with hMm where HiRISE coverage permits observation.Often occurs in slightly lower, rougher areas than adjacent lBg1.The more eroded fractures have more dark infill.Below the resolution of CRISM observations.There is a gradational boundary with lBg1 but a distinct topographic step (∼50 cm) between the two where iDm is exposed.Seen throughout the wider Oxia Planum where scarps or windows though lBg allow.

Lower bedrock group lBg1 Lower
(Continued ) JOURNAL OF MAPS unit exist.Significant differences in expression include variations in fracturing and layering seen throughout the group (lBg1 -lBg3) and Brighter materials (lBg2) having a larger amount of dark regolith trapped in their 'knobby' texture.Only in a few instances are these textural variations clear and it remains an open question as to whether they represent stratigraphically distinct lithologies or lateral variation in preservation and erosion processes, or both.Nevertheless, these mapped contacts are locations to investigate potential lithological variation -, even if the actual significance of these divisions is currently unclear.
In the upper bedrock group (uBg) we also map three units.The uBg dose not have such a strong association with aqueous alteration, but in the wider Oxia Planum region do contain exhumed fluvial channel bodies (Davis et al., 2023;Fawdon et al., 2022).The top of the group (uBg1) forms a thin resistant layer directly underlying both the Mound material group and the overlying Dark material unit.The thin resistant layer is separated from the rest of the group to document the extent of this possible paleo-surface and because the other units (uBg2,3) host a boxwork of upstanding ridges (H) which we interpret to probably be inverted mineralised fractures, which in turn could represent phases groundwater activity of great interest to the mission (e.g.; Kronyak et al., 2019).
We map two units in the Mound material (Mm) group.These units constitute isolated hills (hMm) that make up the lithologically distinct top parts of geomorphological identifiable mounds.These landforms are thought to be remnants of a ∼100 m thick layer (McNeil et al., 2022).However, the deposition and erosion mechanisms for the unit, which may be significant for the preservation of biosignatures, are unresolved.Ridge mound material (rMm) is included separately because it was not always clear if these small outcrops, which appear similar to the top of the mounds, always have a contact at their base or are ridges in other units and an artefact of illumination.
The Dark material (Dm) in the mapping area are found to occurred in two circumstances and are mapped as two units.At the top of the stratigraphy the overlying dark material (oDm) unit is thin (∼<5 m), is rough in textured, contains abundant trapped loose material trapped in its surface.oDm occurs at on local topographic highs (such as inverted craters and channels), or in regional topographic lows surrounded by a shallow moat meaning that it is still at the top of the stratigraphy.Regionally dark material in these settings are associated with fluvial systems (Fawdon et al., 2022) and mafic spectral signatures (Quantin-Nataf et al., 2021).However, a type of dark material is also found within erosional windows and appears to be interbedded with the lBg.Dark material in this setting often appears to be dark sand around bright blocks but the reason for this roughness with Note: Description of map units organised from youngest to oldest as on the Main Map.respect to the low relief surrounding units is unknown.This 'Dark material' would substantially predate the overlying Dark material (oDm) thus we discriminate to aid future investigation identifying interbedded Dark material (iDm) units.Light-toned patches with concentric layers occur in two distinct settings.Concentric bright patches (cBp) occur in small (<100 m) relic impact structures predominantly overlying the lower bedrock group (lBg1).Whilst these may be contemporaneous with the base of the upper bedrock group (Ubg3), their distinct setting and appearance suggests a localised process warranting their distribution be mapped separately.Crater infill bright patches (cfBp) appear similar to cBp but occur within larger (>250 m) craters with well defined rims.Consequently, this crater fill (cfBp) material must substantially post-date the craters themselves and when the upper bedrock group in the craters formed.However, all these light toned patches may have formed at the same time if cBp represents erosional remnants of a process that is simply better preserved as cfBb in the larger craters.These units are mapped separately to avoid interpretive stratigraphic assumptions.
Impact Crater materials (Cm) have been mapped based on three classes of degradation.Craters with the best preserved morphology (nCm) are recorded to identify relatively recent impacts that have occurred, and exposed fresh bedrock.Other crater materials are divided into recent (rCm), where there is not significant reworking of the crater or ejecta, and degraded examples (dCm) that we consider to be older.This class (dCm) records the parts of impact structures associated with impact lithologies (ejecta remnants, disaggregated rims) but other bedrock lithologies, exposed by the impact process, when also visible, are mapped with those stratigraphically appropriate units.

Surficial units
Five surficial units and two texture overlays represent units that crop out discontinuously across the region, deposits of mobile material, and surface patterns that transcend several other units.The Dark (D) and Smooth (S) surficial units are found intermittently across the mapping area.The Dark unit is notable in that it often exhibits a 'stepped' edge, which indicates a degree of induration that is not seen in the Smooth unit (S), which in turn represents what is likely to be unconsolidated material, such as scree at the foot of scarps.Boulders (B) and two classes of aeolian bedforms (T and t) were mapped based mainly on their identification in the NOAH-H data set (Barrett et al., 2023).These units were mapped to help understand the modern aeolian environment (E. A. Favaro et al., 2021) and also because understanding the distribution of these bedforms is vital when considering rover locomotion for long term planning.Finally, the Honeycomb (H) and Fracture (F) textures observed across several units representing specific processes and events associated with their geological history.These are recorded to indicate their distribution for future study, as they may be related to possible paleo-surfaces, regional stress fields, or host resistant fracturefills indicative of ground water activity and may be therefore relevant to the objectives of the rover mission (e.g.Vago et al., 2017).

Structural observations
Three classes of structural observation are recorded on the map.Three probable reverse fault zones are inferred in the Pannonia region.This inference is drawn from the pattern of repeated UGB and LBG units, the topographic break in slope and that the spacing of these features is consistent with the strike scale and spacing of shortening structures in the region (e.g.; Woodley et al., 2023).The 54 lineaments represent a variety of linear features including fractures, 10's-100's m long which the stratigraphy.These may represent faults, but no clear evidence of displacement is seen.We also observed polygonal to rectilinear fractures through to relate to the region tectonic fabric (e.g.; Apuzzo et al., 2021).These recorded as a textural overlay and described with the units in which they occur (Table 4).

Relationship to previous mapping
Previous mapping efforts have focused on many different aspects of Oxia Planum (Figure 4) with regional studies (Ivanov et al., 2020;Tanaka et al., 2014), spectroscopic studies (Brossier et al., 2022;Gary-Bicas & Rogers, 2021;Mandon et al., 2021;Parkes-Bowen et al., 2022), and terrain characterisation works (Mastropietro, 2019) investigating the geological history (Quantin-Nataf et al., 2021;Uthus, 2020) and the proposed landing area.Where these maps overlap the area of this work, a variety of 'clay bearing' or 'light toned' plains units (e.g.CBP) are identified cropping out over the majority of the area mapped, as well as instances of a 'dark capping or resistant units' (e.g.DRU) and examples of 'isolated mounds' are mapped.
Exploration of the CBP has used hyperspectral (CRISM; Brossier et al., 2022;Mandon et al., 2021) and multispectral (THEMIS; Gary-Bicas & Rogers, 2021;CaSSIS;Parkes-Bowen et al., 2022) data supported by HiRISE.Mandon et al. (2021) identified 'Orange' and 'Blue' spectral classes, with their distribution and relationship to fracture patterns mapped by Parkes-Bowen et al. (2022).These classes loosely relate to units in our upper Bedrock (blue, uBg) and lower Bedrock (orange, lBg) groups, but should never be taken as a direct correspondence because our mapping used much higher resolution data and considers geological context independently of interpretation of mineralogical inferences from spectroscopy.Other mapping efforts either did not attempt to differentiate the CBP (e.g.Ivanov et al., 2020;Mastropietro, 2019;Uthus, 2020) at their mapping scale, or the authors divide the CBP based on an erroneous observation of an apparent albedo difference made using the beta release of the Murray lab global CTX mosaic that is, in actuality, a mosaicking artefact with no geological significancelater corrected in the full data release (The Murry Lab, 2023).
There is a good correlation between the DRU of previous works and the Overlying Dark Material (oDm) units.This map and these works did not have sufficient resolution to record what we identify as the interbedded Dark material or that the DRU is only a very thin (<5 m) layer and recessive compared to the subjacent material (uBg3) which forms the cliff edges of the mesas.There is also a direct correlation between the 'mounds' of Quantin-Nataf et al. (2021), the Mound Material group from this study, and a detailed regional investigation of these landforms (e.g.McNeil et al., 2021).

Anticipated use in rover operations
The purpose of this map is to inform the ExoMars rover mission.The map creates a connection between planning for the rover's traverse and regional geological context and we anticipate that it will be used throughout the envisaged reference surface mission (Vago et al., 2017).
After Entry Descent and Landing (EDL), the rover will carry out a 15-day commissioning phase which will include rover deployment, egress, moving away from the lander, and commissioning the rover subsystems, the instruments, and the rover's analytical laboratory chemical background.During this time, we will conduct a validation exercise, connecting the initial PanCam panorama to our mapped geological units.After considering this 'calibration' we will use the map to inform the initial direction taken by the rover.In the case that the landing site is not within the mapped area, then a rapid reconnaissance mapping protocol (Section 2.2) would be implemented as successfully tested for an ExoMars like mission scenario by Balme et al. (2019).
Following commissioning, the mission will consist of a number of traverses and the exploration of various locales.Six experiment cycles (each including drives, identification of prime targets and near surface plus subsurface samplings) and two deep vertical surveys (where at a single location samples are obtained and analysed at 50 cm depth intervals) are planned for the nominal mission.To support this, we use the map to establish a suite of visual hypotheses; that is, the map creates a framework showing us where we can go to investigate different aspects of the regional geology, and what types of questions can be answered at each location.For example, the map shows us where we can ask 'what is the lithology of the upper bedrock group', and then provides a guide for where to move the rover to investigate follow up questions such as 'does the lithology of the upper Bedrock group become more conducive to the preservation of organic material as we go down section?' The map will connect the rover observations made in these locales with our understanding of the spatial and stratigraphic distribution of geological units and allow us to gradually construct a thorough understanding of the paleoenvironment.This understanding of geological context is critical to correctly choose where to conduct experiment cycles and especially the resource-hungry vertical surveys.The end goal is to maximise our chances of collecting samples that may possess diagnostic chemical biomarkers in a good state of preservation.
Finally, the map will be used for long term strategic planning and team communication.If the mapped unit prove to be meaningful, we anticipate that that long terms plan would attempt to explore, or sample, each of the major bedrock groups and detailed hypotheses relating to each of these groups will subject of future companion publications.By making the map available though MMGIS (Calef et al., 2017) planned long-term drive targets, strategically planned traverse routes, and the daily location of the Rosalind Franklin can be displayed together.In MMGIS, the wider team will be able to interact with the strategic decision-making progress and appreciate how the rover sits within the geologic context.By enabling team access to this information, strategic decision-making will be more robust, inclusive, and should lead to more efficient and productive mission progress.

Outlook
This map represents a key part of the ongoing study of Oxia Planum in preparation for the ExoMars rover mission.In addition to usage outlined above the units and observations are applicable to the Oxia Planum region andusing observation from the region beyond the mapping area, are the basis for more general investigation into the geological history of the region.This will identify of hypotheses for the origin of each units that can be evaluated by Rosalind Franklin.The map and associated data will be used in detailed studies.Such as semi-autonomous terrain classification in the Oxia region or quantitative evaluation of 3D geometry of the mapped units.Finaly this map will be a guide for spectroscopic studies ground truthing CRISM hyper-and CaSSIS multi-spectral data with the PanCam (Coates et al., 2017;Cousins et al., 2012), Enfys (Gunn, 2023), MaMISS (De Sanctis et al., 2017) and µOMEGA (Bibring et al., 2017) instruments on the ExoMars rover and also serve as example for future mission planning.

Conclusions
We have created a high resolution geological/morphostratigraphic map of the ExoMars Rosalind Franklin rover landing site in Oxia Planum.The map is presented at a scale of 1:30,000 but digitisation was performed at a scale of 1:2000, a scale relevant to rover operations.Digital version of the map are available (Fawdon et al., 2023).To create the map, we used a two-phase process of group mapping followed by expert reconciliation.The mapping exercise has provided the wider ExoMars rover team with a sound knowledge of the landing site and has also helped us to develop new geological hypotheses for the region.As a summary of spatially located geological hypotheses, the map will be used to guide rover exploration and, when these hypotheses are tested with surface observations, will enable the team to identify suitable locations for sample acquisition using the rover's drill system, and hence improve the chances of the mission meeting its search for life goals.In addition, the map provides the geological framework necessary to provide context for these samples, such that any important discoveries can be tied into the Mars' regional and global geological record.

Software
The map and other datasets were created and compiled in ESRI ArcPro 2.9-3.1.Individual HiRISE Digital Elevation Models were created in SocetSet® and Integrated Software for Images and Spectrometers (ISIS3).Creation of the CaSSIS RGB products was completed using ISIS3.Georeferencing was conducted in ESRI ArcPro 2.9-3.1.The HiRISE orthomosaic and DEM basemap (Volat & Quantan-Nataf, 2020) was produced using the MarsSI infrastructure (Quantin-Nataf et al., 2018 The data collection in the concurrent team mapping used Multi-Mission Geographic Information System (Calef & Soliman 2019).

Figure 1 .
Figure 1.Location.The location of the high-resolution map of the landing site in Oxia Planum, Mars (A) Topographic map showing Oxia Planum on the northern margin of Arabia Terra.White lines mark NW-SE extent.(B) Topographic map (CTX DTM from Fawdon et al., 2021) overlain on THEMIS mosaic showing the extent of the mapping area within Oxia Planum.The 1 and 3 sigma landing ellipses are show in yellow.The different ellipse azimuths represent the start (high azimuth -LPO) and close (low azimuth -LPC) of the planned launch window, with many more ellipses (not shown) aligned between these extremes reflecting launch at any time within the launch window.
in local topographic lows.Patches are progressively more dissected from SE-NW.Regionally they are found close to recent impact craters.F Widely spaced fractures -HiRISE: Polygonal to rectilinear troughs ∼1-5 m wide with spacings of 10 m -20 m.CaSSIS: Dark lineations in other units.
Overlying dark material -HiRISE: Dark-toned, rough, boulder-forming ridges with dark smooth troughs.CaSSIS: Predominantly blue tones with small, dark blue zones.Located on local topographic highs.The unit's margin is recessed from the tops of scarps and it is sufficiently thin (∼ 1 m) to contain occasional windows into underlying lighter toned units.CRISM pyroxene signatures(Quantin-Nataf et al., 2021) and possible weak phyllosilicate signatures.Appears to unconformably overlie F in unit uBg1.Found in regional topographic lows, on local topographic highs, often surrounded by a moat.Often found in continuum with inverted and low relief channels(Davis et al., 2023;Fawdon et al., 2022).iDmInterbeddeddark material -HiRISE: High relief, bright ∼1-3 m diameter rounded knobs picked out by dark infilling material.Occasional bright blocks.CaSSIS: Dark or mottled because the bright knobs are often below the CaSSIS resolution, whereas the unit is lower relief, bright purple-orange tones.Located in local topographic lows.Appears to crop out in erosional windows within bedrock units.No phyllosilicate signatures observed.This unit appears to occur sporadically and be interbedded with lBg2.Similar examples of this material are seen occasionally across the region.Bright group cfBp Crater fill bright patches -HiRISE: Mid to bright toned, low relief with concentric dark lineation.CaSSIS: Bright/white tones.Located within degraded impact craters > ∼150 m.No observed compositional relationships.Strongly associated with dCm.Overlain by surficial units.Not seen to be overlain by other units.Common throughout the region in degraded impact craters > 150 m in size.It is not seen to be overlain by anything other than surficial units.cBp Concentric bright patches -HiRISE: Mid to bright toned, low relief with distinct concentric dark lineation.CaSSIS: Bright/white tones contrasting to subjacent bluer tones.Located in local topographic lows with distinct marginal steps No observed compositional relationships.Overlies lBg units.The moat around the edge often hosts aeolian material.Not seen in association with uBg units and occasionally overlain by rMm.Common thoughout the region where the orange toned material is exposed.Regional observations suggest cBp is superposed by material from the uBg, and Mm unit groups (McNeil et al., 2022).Mound group hMm Hill mound material -HiRISE: Smooth, bright, conical mounds ∼100 m across and up to 10 m in height.The bright smooth surface has occasional lineations.CaSSIS: Yellow tones in all illumination angles.hMm forms the tops of larger mounds.Mounds often have a bright layer half way up, scree around their bases and occasionally have small impacts on them.No good CRISM observations are avaliable in Oxia Planum.Directly overlies uBg1 with uBg2 and uBg3 forming the base of the topographic mound below hMm.hMm never overlies oDm.Mounds with hMm are part of a population extending 100s of km into Chryse Planitia(McNeil et al., 2022).In Oxia Planum hMm must predate oDm.Some mounds are dissected by troughs and have internal ridges.

lBg3
Lower Bedrock Group 3 -HiRISE: Bright, low relief surface with moderately (∼5-10 m) spaced fractures that show no strong orientation pattern.CaSSIS: Low relief with bright, light orange tone.Only seen at the lowest elevations of the map (Pannonia and impact craters).The margins can appear to grade into the other surrounding units.Very strong spatial correlation with phyllosilicate spectral signatures.It could underlie or be at the same level as examples of iDm.Elsewhere in the region this unit often occurs in areas with bright orange tone in CaSSIS data, and occur in the lowest parts of the topography.

Figure 4 .
Figure 4. Relationship to previous mapping.A comparison between this and previously published maps covering Oxia Planum.(a) Simplified version of this map.(b) Context from the global geological map showing that our study area lies in a Late-Noachian highland (lNh) unit (for other labels see Tanaka et al., 2014).(c) Preliminary mapping from Quantin-Nataf et al. (2021) established the key regional units including the 'Clay-Bearing Plains' (CBP) and 'Dark Capping Units' (DCU) and the mounds.(d) Uthus (2020) mapped the CBP as one unit, but (e) Ivanov et al. (2020) and (f) Mastropietro (2019) divided this unit in the middle of our study area.Inspection of this boundary shows that it is a mosaicking artefact seam lines from The Murry Lab (2023) Beta release (shown in red), not a real geological boundary.(g) Mandon et al. (2021) split the CBP into different classes using HiRISE and CRISM to identify 'orange' and 'blue' sub-units, the distribution of which are shown mapped in (h) by Parkes-Bowen et al. (2022).For full description of each map see the respective publications.

Table 2 .
CaSSIS data used observations.
Note: CaSSIS observations used for mapping during the reconciliation phase shown in Figure2(d).

Table 4 .
Description of map units.
(Berman et al., 2018)and impact craters(nCm,  rCm).There are no resolvable compositional associations.Overlies all other units.Common throughout the region in similar contexts associated with Dark material and recent impact craters.Overlies other units and are closely associated with units T and S. Common throughout the region particularly in lower elevations towards Chryse Planitia.T Large TARs -HiRISE: Bright faced elongated ridges.CaSSIS: High relief, high albedo ridges with orange tones on illuminated slopes.Often found in topographic lows (e.g.crater floors) or channels.Ridges are resolvable in DEM data.No clear compositional signature(Berman et al., 2018).Overlies other units as isolated TAR's or grades into smaller continuous, often lower relief, TAR fields (e.g.t).Found in similar contexts throughout the region and strongly associated with sediment fan remnants in Ratia Palus.(Continued)JOURNAL OF MAPS
Overlies other units and closely associated with unit t.Common throughout the regions where loose material accumulates in local topographic lows.
Knobbles and ridges both overlie uBg and lBg group.Similar ridges are frequently seen thoughout the wider Oxia Planum region commonly trend into PBRs (Favaro et al., 2021).Overlies uBg2 It is overlain by examples of both the Mm and Dm groups.Similar units appear at the top of scarps and on local topographic highs throughout the region where they are often overlain by Mm and Dm group style units.